Who needs to learn AI? Debunking a myth with Craig Clawson, Director at the NVIDIA Deep Learning Institute

The World Economic Forum has recently updated its forecast on the new jobs that will be created in the next five years as a result of the technology revolution to a staggering 97 million.

In this context, and with the pandemic accelerating the online shift of a large part of our daily lives, it is clear how widespread digital skills are not a ‘nice-to-have’ anymore – if this was ever the case.

We sat down with Craig Clawson, Director at the NVIDIA Deep Learning Institute, to talk about misconceptions in artificial intelligence (AI) learning and digital skills and their importance for society, and to take stock of where Europe stands in the digital education and vocational training landscape.

Both DIGITALEUROPE and NVIDIA are partners of the EU’s Digital Skills and Jobs Coalition, which is tackling the digital skills gap through a range of initiatives – including AI learning – aimed at students, small businesses, and ICT professionals all over Europe. DIGITALEUROPE is now helping the Coalition create a one-stop-shop platform for digital skills and jobs in Europe to support National Coalitions and better exchange information on best practices, funding and training opportunities, and EU initiatives.

Craig has a fascinating background: while not a computer scientist by training, he specialises in helping individuals and companies close this digital skill gap and is perhaps the perfect example of how learning about AI can change your career.

Craig Clawson, Director, NVIDIA Deep Learning Institute

Hi Craig, and thanks for having a chat with us. Let’s see if we can start by debunking a myth. Is learning AI only for data and computer scientists?

AI is already touching our everyday lives. It’s used by air traffic controllers to keep us safe when we fly. It assists doctors in diagnosing diseases by powering machines that can identify and understand images. AI helps prevent fraud in banking and, in agriculture, it’s helping improve crop yields and enable more sustainable farming. It’s hard to think of an industry that hasn’t already or won’t be impacted by AI in the future.

Of course, we don’t all need to be computer or data scientists to gain value from interacting with AI. Even if we are not developing AI applications ourselves, by embracing AI in our everyday lives, we can also help it to learn and become even more impactful. For those making career decisions, incorporating an understanding of AI and its applications in a particular field is a wise move.

The World Economic Forum’s 2018 report predicted that there will be 58 million new jobs in AI by 2022. While the number of education or training programs in this area is rapidly increasing, those with the necessary skillset are still hard to come by.

Not all of these roles will require computer science or engineering degrees. In fact, someone who can combine expertise in an application-specific field with some understanding of data and machine learning will be especially valued. For instance, a sought-after AI healthcare specialist would combine an understanding of data analytics with a background in human biology.

How about your own background and career? You’re not a computer scientist, yet you deal with AI and accelerated computing. Do you see digital technologies intersect more and more with non-technology career paths in the future?

I began my career thinking I would be an academic economist. When I was in graduate school, I was drawn to a developing field called behavioural economics. I never imagined how much AI and the exponential increase in data available to social scientists would enhance research in this discipline – and, conversely, how behavioural economics is now impacting the work data scientists do.

So the intersection you describe is very exciting for me. I came to my role at NVIDIA with a limited understanding of accelerated computing and AI, and now it is my responsibility to help others learn these technologies. The application and use of AI is advancing at an incredibly fast pace, and it is thrilling to have a front row seat to this technology that is changing the world around us.

I think those who have taken non-technological career paths but are open to learning about how technology is shaping the world, will have an especially important role in our society. We will need these individuals to map out the domains for AI and provide input into how applications can best serve society. Designing systems whereby humans interact and work with AI-driven technologies will not be the sole province of data scientists.

“We don’t all need to be computer or data scientists to gain value from interacting with AI. I think those who have taken non-technological career paths, but are open to learning about how technology is shaping the world, will have an especially important role in our society.”

What’s the story behind the NVIDIA Deep Learning Institute? And what did you achieve to date, in Europe and globally?

The NVIDIA Deep Learning Institute (DLI) was established to train developers, data scientists, and researchers in how to use deep learning and accelerated computing. Since DLI began, we’ve trained over 200,000 developers globally, with around 20% of those in Europe.

We offer hands-on training that focuses on solving real-world problems and developing actual applications with deep learning frameworks. Our courses cover AI applications such as robotics, self-driving cars, natural language processing, and more. More recently, we’ve expanded our training offering to other domains, including data science and high-performance computing, as well as providing courses for IT professionals. For their training, every student gets access to a development system in the cloud accelerated by NVIDIA graphics processing units (GPUs).

We want to make AI training accessible to the entire developer ecosystem and therefore partner with a number of public and private organizations to reach developers who are new to AI. Universities and higher education institutions can offer our training courses for free through the use of NVIDIA Teaching Kits and our DLI University Ambassador program. The former provides qualified university educators access to GPU cloud resources and hands-on course materials across deep learning, accelerated computing, and robotics, allowing them to integrate these into their existing curricula.

Several universities in Europe are already taking advantage of this opportunity. The DLI University Ambassador program provides free workshops for students led by NVIDIA-certified ambassadors. We already have close to a hundred ambassadors across European institutions. Our workshops are also part of the Partnership for Advanced Computing in Europe (PRACE) training program thanks to our collaboration with the ambassadors at the Leibniz Supercomputing Center and IT4I in the Czech Republic, which help bring AI skills to the HPC community.

Our membership in DIGITALEUROPE is key for us in driving AI upskilling initiatives in the region. We have recently pledged to provide 200 free self-paced online courses to advance the AI skills of developers and data scientists at small and medium-sized enterprises.

Beyond DLI, NVIDIA has a rich developer program that provides free access to forums, software and technical documentation to make getting started with GPU projects easier. Our digital GPU Technology Conference (GTC) delivers hundreds of talks on AI and accelerated computing, and is free of charge for university, government, and non-profit attendees.

Through the Digital Skills and Jobs Coalition, DIGITALEUROPE and NVIDIA are working to bridge the AI skills gap through education. Why is this so important? What can the EU do to close this skills gap, now that the issue is at the top of the agenda?

As a man-made tool, AI can be used for a broad range of purposes – the vast majority will improve productivity, products and services, and benefit humanity. AI will create dislocations but history has shown that, over time, new technology is the engine of economic growth. For one, AI will augment many jobs – helping doctors, lawyers, truck drivers work more efficiently. It will enhance productivity, as the number of accidents, trauma calls, ambulance calls and insurance claims decrease. Third, it will drive costs down and enhance quality of life. I believe that AI will make us better and more productive in the work we do and will actually make our careers more secure, not less. 

We see an increasing number of governments around the world taking an active role in building out their respective high-performance computing capabilities and providing opportunities for their people to develop necessary skills. NVIDIA collaborates with national governments through its AI Nations initiative, a worldwide program that helps leaders and stakeholders develop plans for implementing AI to drive economic growth and advance national priorities. Helping countries and organizations close the skill gap is something we are passionate about and involved in regularly. To address the issue in Europe, the EU should continue focusing on initiatives including fostering explicit AI strategies and plans, creating communities of interest – such as the Digital Skills and Job Coalition –, sponsoring AI pilots in the public sector, nurturing AI startup ecosystems, and forging connections between government, industry and academia.

“Our membership in DIGITALEUROPE is key for us in driving AI upskilling initiatives in the region. We have recently pledged to provide 200 free self-paced online courses to advance the AI skills of developers and data scientists at small and medium-sized enterprises.”

One of the key lessons of the pandemic is that no one can afford to overlook their digital skills anymore. How about smaller businesses? Why is learning AI important for them?

The world has been through a lot this year and hopefully, together, we’ve all learned something of value. Many of us have had to adjust to working from home and connecting with colleagues and peers remotely. Certainly digital skills have enabled many of us to continue on in our regular work, hardly missing a beat. For many others though, their work may have grinded to a halt or suffered fits and starts as governments have oscillated between opening and closing certain sectors in their economies.

Organisations of every size are looking at how investments in technology can increase value and enable them to operate more efficiently in this new world. A recent Bain & Company study found that three out of four companies plan to accelerate automation initiatives due to COVID-19.

Businesses are adopting certain technologies so they can be better prepared if the current situation drags on or for the next crisis. Even small businesses can benefit from adopting some of these technologies.

Recently I read about how relatively simple simulations and algorithms were used to help a small coffee shop chain predict and mitigate effects on the reliability and pricing of coffee supplies due to weather changes and natural disasters.

This is just one example. Customer-relationship management (CRM) systems are becoming infused with AI applications that can help improve sales and marketing. AI tools can help facilitate better hiring and onboarding processes. And customer communications can be enhanced, for example through the use of a chatbot. There is no doubt that AI tools are becoming increasingly accessible to small and medium-sized businesses.

In the future, do you think AI knowledge will become a horizontal competence rather than a specialized skill? Or will we rather move towards software which doesn’t require in-depth knowledge to be used?

This is an excellent question – it really highlights the varying ways AI applications can be developed and deployed.

The answer that I think is best for this is probably the least satisfying – and that is I can see both models – depending entirely on the software application and its purpose. A simple and easy user interface is key to any AI application intended for a mass audience. Those applications will likely require minimal technical knowledge on the user’s part, and they may even use AI to monitor, update and repair themselves. However, applications that offer greater flexibility, involve greater complexity in defining and developing use cases, or require more human input for design and ongoing deployment, will likely require competencies across varied skill sets. They will require more insight and input than data and computer scientists can provide alone.

You can find more information about NVIDIA’s pledge to the Digital Skills and Jobs Coalition here. For inquiries around NVIDIA’s training programs in Europe, please reach out to Mrs. Marjut Dieringer, Senior Manager Deep Learning Institute EMEA, at mdieringer@nvidia.com.